physiclaw/PhysiClaw

The AI agent that physically operates a phone — the way you do.

What it solves

PhysiClaw is designed to handle everyday phone errands—such as ordering food, shopping for groceries, or paying bills—without needing app APIs, OAuth, or software-based automation tools. It bypasses anti-bot detection systems by interacting with the phone physically rather than through software simulations.

How it works

The system uses a camera to watch the phone screen and a robotic stylus to perform taps and gestures. It operates as an AI agent that you can communicate with via WhatsApp; you send it a request in plain language, and the agent wakes up, unlocks the phone, performs the task, and notifies you of the result.

Who it’s for

Users who want to automate tasks across various mobile apps that lack public APIs or have strict anti-bot measures, and those who have the required hardware (robotic arm, stylus, and camera) and an iPhone.

Highlights

  • Physical interaction: Uses a camera and stylus to mimic human finger touches, making it indistinguishable from a real user to the phone.
  • API-less automation: Works with any app on the phone without needing ADB cables or software installations.
  • Natural language interface: Controlled via WhatsApp messages.
  • Open source: Both the agent software and theCAD-as-code hardware designs are provided under the MIT license.

Related

  • Project
    ModalityDance/PalmClaw

    An AI agent framework that runs natively on Android, managing sessions and tool execution on-device while using external LLM APIs for inference.

  • Project
    minitap-ai/mobile-use

    An open-source AI agent that automates Android and iOS devices using natural language to perform tasks and scrape data from apps.

  • Project
    kellyvv/PhoneClaw

    A local AI agent framework for iOS that runs LLMs on-device to privately manage calendar, health data, and device tasks via natural language.

  • Project
    yashab-cyber/opendroid

    An autonomous AI agent for Android that plans and executes multi-step tasks across various apps using UI automation and a wide range of LLM providers.

  • Project
    ttguy0707/CyberClaw

    A LangGraph-based general-purpose AI agent framework that implements a secure two-phase skill execution protocol and sub-agent delegation for observable and extensible workflows.